"Pete C." wrote:
>
> Mark Thorson wrote:
> >
> > This study gives numbers comparing farmed vs. wild,
> > and there's a huge difference, especially at the
> > high end of the range.
> >
> > Environ Sci Technol. 2006 Sep 1;40(17):5347-54.
> > PCBs, PCDD/Fs, and organochlorine pesticides in farmed
> > Atlantic salmon from Maine, eastern Canada, and Norway,
> > and wild salmon from Alaska.
> > Shaw SD, Brenner D, Berger ML, Carpenter DO, Hong CS,
> > Kannan K.
> > Marine Environmental Research Institute, P.O. Box 1652,
> > Blue Hill, Maine 04614, USA.
> >
> > Farmed Atlantic salmon (Salmo salar) from Maine and
> > eastern Canada, wild Alaskan Chinook salmon (Oncorhynchus
> > tshawytscha), and organically farmed Norwegian salmon
> > samples were analyzed for the presence of polychlorinated
> > biphenyls (PCBs), dioxin-like PCBs, polychlorinated
> > dibenzo-p-dioxins (PCDDs), dibenzo-p-furans (PCDFs), and
> > chlorinated pesticides. PCDD and PCDF congeners were not
> > detected in > 80% of the samples analyzed. Total PCB
> > concentrations (7.2-29.5 ng/g, wet weight, ww) in the
> > farmed salmon were significantly higher than those in
> > the wild Alaskan Chinook samples (3.9-8.1 ng/g, ww).
> > Concentrations of PCBs, WHO PCB TEQs, and chlorinated
> > pesticides varied significantly by region. PCB and WHO
> > PCB TEQ concentrations in farmed salmon from eastern
> > Canada were lower than those reported in samples
> > collected two years earlier, possibly reflecting recent
> > industry efforts to lower contaminant concentrations in
> > feed. Organically farmed Norwegian salmon had the highest
> > concentrations of PCBs (mean: 27 ng/g, ww) and WHO PCB
> > TEQs (2.85 pg/g,ww); their TEQ values are in the higher
> > range of those reported in farmed salmon from around the
> > world. Removal of skin from salmon fillets resulted in
> > highly variable reductions of lipids and contaminants,
> > and in some skin-off samples, contaminant levels were
> > higher, suggesting that skin removal does not protect
> > the consumer from health risks associated with
> > consumption of farmed salmon.
>
> What I see from that data is that levels in both are quite variable, and
> indeed some of the farmed samples had lower levels (7.2 ng/g) than some
> of the wild samples (8.1 ng/g).
>
> Showing just the high and low of a group of samples is misleading since
> you don't know what the distribution is. Those extremes may represent
> only one sample, and make it very easy to distort the data to fit an
> agenda. The only mean value given was for the organic farmed Norwegian
> samples.
>
> It's entirely possible that most of the samples from both groups were in
> the 7-8 ng/g range showing no difference between farmed and wild, or
> most of the wild near 4 ng/g and most of the farmed near 29 ng/g showing
> a huge difference.
>
> Again, it's incomplete data and subject to presentation in a manner that
> may be misleading. Either way we're talking about extremely low levels.

It's an abstract. The authors' conclusions are given
in the abstract, but the detailed data is in the full
paper, which is not online.

But no competent scientist would make an assertion
in an abstract that wasn't backed up by their data.

Environ Sci Technol. 2005 Oct 1;39(19):7389-95.
Polychlorinated biphenyls in salmon and salmon feed:
global differences and bioaccumulation.
Carlson DL, Hites RA.
School of Public and Environmental Affairs, Indiana
University, Bloomington, Indiana 47405, USA.

Concentrations of 160 polychlorinated biphenyl (PCB)
congeners or congener groups were determined in approximately
600 farmed Atlantic salmon from around the world and wild
(ocean-caught) Pacific salmon from the Northeast Pacific.
Concentrations and PCB congener profiles were analyzed
to provide insight into the sources and uptake of PCBs in
salmon as well as regional differences. Although total PCB
concentrations in wild salmon appeared to be correlated to
total lipid content, the increased proportion of total lipids
in the farmed salmon could not account for the much greater
PCB concentrations. We investigated the PCB congener patterns
of hundreds of salmon samples using principal component analysis
to further illuminate regional and species differences. Three
major PCB patterns were observed, in most wild fish (except
British Columbia and Oregon chinook), in farmed fish from the
Atlantic, and in most farmed fish from the Pacific. The PCB
congener profiles of farmed salmon often closely corresponded
to a sample of commercial feed purchased in the same region,
indicating that the feed is likely to be the major source of
PCBs for farmed salmon. In such cases where PCB profiles in fish
and feed were similar, a comparison of congener concentrations
in fish and the feed showed that the majority of congeners, with
some notable exceptions, were bioaccumulative to the same extent,
irrespective of physical properties.